With the rapid development of production and technology, the control system automation level has been raised continuously, and the scale of control system has become more and more large. However, because the complexity of the system has been raised and the systems work under high-power and high workload in long time, the component failures are almost inevitable. These failures can cause a serious decline in the performance of the system and even cause the out of control. As a control method that can improve the stability of complex system, fault-tolerant control emerges.The technology of passive fault-tolerant control is to design controllers with fixed structure. So that, both in the normal operating conditions and in the condition actuators, sensors and other parts fail, the controllers can guarantee the system stability and some satisfactory performance. Because there no exists the unit of fault detection and isolation in the passive fault-tolerant control system, there is few time-delay and it becomes easy to design the passive fault-tolerant control system.In this thesis, a detailed analysis is stressed about the passive fault-tolerant controllers of linear MIMO systems and design procedures of controllers. Since in some real application systems, sensors and actuators are components that fail most frequently and are the most important, hence, in this thesis, failures in actuators are mainly considered. A break-normal failure model is used in the thesis to represent the seizure and normal mode of actuators. Based on linear matrix inequalities (LMIS), a sufficient condition of the existence of passive fault-tolerant controllers and the solution are given.The several problems studied in this thesis are as follows based on LMI:1. The design of the passive fault-tolerant control system of linear time-invariant systems and systems with uncertainty is investigated based on linear matrix inequality (LMI) approach. The sufficient condition is given to obtain the asymptotic stability of the linear system and system with uncertainty, considering actuator failure using state-feedback methods. With this, the fault-tolerant controller is designed to apply the conditions.2. In some real application systems, uncertainty parameters and time-delay are inevitable. Therefore, in this thesis, the design of robust fault-tolerant controller of the uncertain system with time-delay is investigated through memory-less state-feedback and memory state-feedback methods based on linear matrix inequality (LMI) approach. The sufficient condition is given to obtain the asymptotic stability of uncertain system with time-delay, based on actuator failures. With this, the robust fault-tolerant controller is designed to apply the conditions.3. In the first two problems, the system stability is mainly considered, while in chapter5of this thesis, either the system stability or the dynamic performance is considered through pole assignment. The design of the passive fault-tolerant controller considering the poles assignment of linear time-invariant systems is investigated based on linear matrix inequality (LMI) approach. The necessary and sufficient conditions for the existence of passive fault-tolerant controller, which enables the poles of the closed loop system with actuators failure to lie in a prescribed region, are proposed. Based on this, a solution to get the fault-tolerant controller is given. |